▸ DEVICE UNDER TEST
NVIDIA CMP 170HX 8 GB — 8 GB VRAM.
▸ CMP 170HX 8 GB SPEC
- BRAND
- NVIDIA
- VRAM
- 8 GB HBM2e
- BANDWIDTH
- 1490 GB/s
- FP16 COMPUTE
- 50.5 TFLOPS
- FP32 COMPUTE
- 12.6 TFLOPS
- CUDA CORES
- 4,480
- TENSOR CORES
- 280
- TDP
- 250 W
- ARCHITECTURE
- Ampere
▸ AI CAPABILITY
220/ 449 models @ Q4
With 8 GB VRAM and 1490 GB/s bandwidth, this GPU handles models up to 10.7B parameters.
Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~170 tok/s.
§ 01TOP MODELS FOR CMP 170HX 8 GB
220 FIT · SHOWING 20| MODEL | SIZE | VRAM Q4 | TOK/S | AVG |
|---|---|---|---|---|
| SOLAR-10.7B | 10.7B | 7.0 GB | 111 | 28.2 |
| Falcon3-10B | 10.3B | 6.8 GB | 116 | 38.2 |
| GLM-4.1V 9B Thinking | 10.29B | 6.8 GB | 116 | — |
| Bamba 9B v2 | 9.78B | 6.5 GB | 122 | 26.1 |
| Qwen 3.5 9B | 9.65B | 6.4 GB | 124 | 50.6 |
| RecurrentGemma 9B | 9.63B | 6.4 GB | 124 | 35.0 |
| glm-4-9b | 9.4B | 6.2 GB | 127 | 20.5 |
| MiniCPM-o 4.5 | 9.37B | 6.2 GB | 127 | — |
| gemma-2-9b | 9.2B | 6.1 GB | 130 | 30.2 |
| Yi 1.5 9B | 9B | 6.0 GB | 132 | 30.3 |
| Yi Coder 9B | 9B | 6.0 GB | 132 | 35.8 |
| Ministral 3 8B | 8.92B | 5.9 GB | 134 | 25.7 |
| Ministral 3 8B Reasoning | 8.92B | 5.9 GB | 134 | — |
| NVIDIA-Nemotron-Nano-9B-v2 | 8.9B | 5.9 GB | 134 | 44.2 |
| InternLM3 8B Instruct | 8.8B | 5.9 GB | 135 | 38.7 |
| Qwen3-VL 8B Instruct | 8.77B | 5.8 GB | 136 | 26.4 |
| MiniCPM-V 4.5 | 8.7B | 5.8 GB | 137 | 26.1 |
| MiniCPM-o 2.6 | 8.67B | 5.8 GB | 137 | 40.3 |
| Gemma 1 7B | 8.54B | 5.7 GB | 140 | 24.7 |
| CodeGemma 7B | 8.54B | 5.7 GB | 140 | 40.2 |